Decoding Gene-Disease Associations with Computational Methods: A Survey

Area of Science:

• Computational biology
• Bioinformatics
• Artificial intelligence in medicine

Background:

• Identifying gene-disease associations (GDAs) is crucial but challenging due to vast data and experimental limitations.
• Computational approaches are essential for prioritizing GDAs and integrating omics data.
• Advances in AI, particularly GNNs and LLMs, offer powerful tools for GDA prediction.

Purpose of the Study:

• To provide a unified and up-to-date overview of AI-driven GDA prediction.
• To summarize data resources, review methodological developments, and discuss applications.
• To highlight challenges and future directions in AI for GDA prediction.

Main Methods:

• Review of public data resources for genes, diseases, and biological information.
• Analysis of methodological advancements from traditional network-based methods to deep learning, GNNs, and LLMs.
• Discussion of AI applications in gene prioritization, drug repurposing, and clinical research.

Main Results:

• AI, especially GNNs and LLMs, enables richer biological representations and more accurate GDA prediction.
• A comprehensive overview of current AI techniques and their practical impact is presented.
• The survey integrates data resources, methods, and applications for robust GDA prediction.

Conclusions:

• AI techniques are transforming GDA prediction, accelerating biological discovery.
• This survey provides a foundation for developing more robust, interpretable, and clinically actionable computational tools.
• Publicly available resources and literature are curated to support future research.